Today the OpenAI API was very stable, I didn't have any timeout so here is a new demo of my latent web browser!
There are a lot of fails, but it is easy to close the offending tab and click on another link to throw the dices again 🎲
(note: video accelerated for brevity)
Conversation
Replying to
The first thing that came to mind was this is a Generative Web.
2
3
18
Replying to
Also a big difference with traditional app making is the level of customization it could reach: instead of passive one-size-fits-all content (of which we are only spectators or users) in the future you could get anything you want immediately, tailored just for you
6
Replying to
Yes!! First time I see someone actually using it in the wild 🎄
2
9
Show replies
Replying to
Is this a web browser that generates the page instead of visiting an existing page?
2
1
17
Replying to
Hello! Yes that’s it!
There are two prompts, one example-based to generate a fantasy list of results in JSON, and one purely instruction-based for the actual HTML+JS+images content
(the prompt is not perfect yet, but I hope we are only at the beginning)
7
Replying to
Do not hesitate to try it for yourself!
Right now this need a bit of work, you need to get the repository and build the app (GitHub.com/jbilcke/web4)
But I will try to see if I can also provide a readymade app for Mac and windows
Let me know if you have any feature request!
1
2
Replying to
Amazing. Where do the images come from? Does it just randomly guess the correct URLs, or do you use some AI generation model?
1
2
Replying to
Thank you! Those are 1024x1024 images generated by DALL·E 2
1
2
Show replies